This training grant is being sought by an MD-PhD student preparing to enter the dissertation phase of her PhD in clinical epidemiology. The candidate aspires to become a leading independent physician-scientist in the field of critical care outcomes research. She will devote 100% of her time to this research for the duration of her PhD. Her hands-on training during her dissertation phase, coupled with carefully selected didactic courses, are designed to make her expert in the use of advanced statistical methods to define associations between critical care exposures and outcomes as well as in the use of qualitative methods to deeply explore mechanisms and generate new hypotheses. This dual training approach will optimally prepare her for future studies on doctor-patient communication during transitions of care for critically ill patients. Throughout this award, she will be mentored by senior experts in health services research, critical care, biostatistics, bioethics, and qualitative methods. The research portion of this career development award will proceed in three complimentary projects to improve understanding of how the circumstances surrounding ICU discharge affect the outcomes for acute respiratory failure (ARF) patients and their surrogates. In project 1, the candidate will examine the effect of ICU census on the medical outcomes of ARF patients using a national database of ICU patients. First, she will evaluate the effect of ICU census on the time between extubation and ICU discharge for ARF patients requiring mechanical ventilation. She will then examine the relationship between ICU census and in-hospital death and ICU readmission, investigating whether time between extubation and ICU discharge acts as an effect modifier between census and negative outcomes. In project 2, she will prospectively collect data to examine psychological and trust outcomes patients who suffered an episode of ARF requiring mechanical ventilation and their surrogates. She will conduct telephone interviews with patients and surrogates one month following ICU discharge and administer validated surveys that assess anxiety, depression, and trust outcomes. This research will therefore produce valid measures of the previously unexplored risks surrounding the time of discharge to ARF patients requiring mechanical ventilation and their surrogates. It will build the foundations of a conceptual model that will be used as the basis for future qualitative studies that examine the content of conversations between the health care team and patients and surrogates at the time of ICU discharge to determine which communication elements patients and surrogates find most helpful. The ultimate goal is to improve outcomes in this vulnerable group of patients and surrogates with future intervention studies targeting bed allocation decisions and communication between the health care team and patients and surrogates.